The Mechanism of the Schmidt Reactions and Observations on the

Huijun Dong, Meihua Shen, Joanne E. Redford, Benjamin J. Stokes, Ashley L. Pumphrey, and Tom G. Driver. Organic Letters 2007 9 (25), 5191-5194...
0 downloads 0 Views 367KB Size
SCHMIDT REACTIONS AND CURTNSREARRANGEMENT

Jan., 1948

317

[CONTRIBUTION FROM THE CHEMISTRY LABOIUTORY OF THEOHIO STATE UNIVERSITY]

The Mechanism of the Schmidt Reactions and Observations 09 the Curtius Rearrangement BY MELVINs. NEWMAN AND HYMAN L.GILD EN HORN^

During studies on the synthesis of 4-amino-2,6dimethylbenzoic acid we made an observation which gave us a clue as to the mechanism of the Schmidt reactions’ and led us to predict that the Curtius rearrangement should be acid catalyzed. In this paper we present mechanisms for the reactions of hydrazoic acid with acids and other compounds and also demonstrate that the Curtius rearrangement is catalyzed by protonic and nonprotonic acids. The original observation which gave the impetus to this work was the fact that 2,6-dimethylterephthalic acid, I, yielded 4-amino-3,5-dimethylbenzoic acid, 11, in the Schmidt reaction. Since the hindered carboxyl group reacted exclusively

Q-@ COOH

CH

\CH,

NHt

Ha04

CHI \CHI

acid is to convert the acid, RCOOH, into an oxocarbonium ion, R-C+=O, or a dihydroxycarbonium ion, R--C+(OH)2.7 The carbonium ion then attacks the nitrogen atom in hydrazoic acid which is attached to hydrogen. The unstable initial adduct loses nitrogen as a molecule of Nz leaving the remaining nitrogen electronically deficient. This in turn sets up forces which cause the R group originally attached to the carbonium-arbon to rearrange to the nitrogen. In this migration the R group carries with it the pair of electrons which originally bound it to the carboniumcarbon. The h a l resulting electron-deficient complex may either remain as such in the reaction medium or, on dilution with water, react with water to yield the final products. 0 R-C

II

+ :N:&:::N: ..-

+ [R4::1T\]’+

H

OOH I

we were struck by the similarity between this Schmidt reaction and the sulfuric acid esterification of 2,4,6-trimethylbenzoic acid.‘ Since the mechanism for this esterification involves an oxocarbonium it seemed logical to propose a similar intermediate for the Schmidt reaction. Accordingly we submit that the Schmidt reaction with acids may proceed by an electrophilic attack of an oxocarbonium ion on nitrogen accompanied by a rearrangement.’ The function of the sulfuric (1) Part of the material herein presented is contained in the M.S. thesis presented by H.L. G. to the Ohio State University in June, 1947. This also formed the basis for part of the addreas given by M. S. N. t o the Organic Symposium in Boston, MossPchulett., June 12, 1947. Additional work on the Schmidt reaction carried out in the Department of Chemistry of the Mossachusetb Institute of Technology by Dr. Conrad Schuerch with Professor E. H. Huntress, executed independently of but concurrently with the present paper, will shortly be presented to THIOJOURNAL. (2) Present address: Cornel1 University M e d i a l School, New York. N. Y. (3) Wolff in Adams “Organic Reactions,” Vol. 3, John Wiley and Sons, Inc., New York, N. Y., 1946. p. 308-310. Under the heading of Schmidt reactions are included the reactions of organic compounds with sulfuric and hydrazoic acids. (4) Newman, THISJOURNAL, 6s. 2431 (1941). (5) As a suitable system for the nomenclature of simple carbonium ion6 we propose that the central electron deficient carbon atom be used as a root and that the attached groups he named as usual. Thus (CHI)IC would be called trimethylcarbonium ion, CHIC *(OH)$ Would he called methyl dihydroxy carbonium ion, and CH-I C*-0, methyl oxo carbonium ion. For more complicated carbonium ions a more systematic nomenclature would be necessary. (6) It is t o be understood throughodt our discussion that the intermediate stages pictured in formulu have no significance 0ththan t o indicate formation and loosening of the bonds involved and t o indicate driving forces for the reactions.

+ Na

[R:::C=O]+

COOH I1

+ CO,

fRNHCOl+ f HzO +’RNHs+ OH

HOC -+

I

A

+ :N:N:::N: ..

H

(1) (la)

--*

1’

+

R N H & O H ) ~+R N H ~ + CO,

(2s)

Evidence supporting the first alternative was obtained by showing that with sulfuric and hydrazoic acids: (A) methyl 2,4,6-trimethylbenzoate, which has been shown to yield (CHa)aC6H*C+=O ions in the cold in sulfuric acid solution,*is converted into mesidine (2,4,6-trimethylaniline) a t 0’ ; (B) 2,4,6-trimethylbenzoic acid which likewise yields (CHa)aCsHaC+=O ions in the cold in sulfuric acid solutions is converted into mesidine, even at 0’ ; (C) benzoic acid, which yields mainly CsHhC+(OH)1ions in the cold in sulfuric acid, does not yield aniline in the cold. This evidence indicates that oxocarboiiiurn ions5 react readily with hydrazoic acid but does not exclude the pos(7) The activated form for hydrazoic acid proposed by Hurd in Gilman’s “Organic Chemistry,” Vol. I, John Wiley and Sons, Inc.. New York, N. Y., 1938, p. 699, is merely one of the contributing resonance structures and does not require aulfuric acid for its existence. Sce Hendriclrr and Pauling, THISJOURNAL, 47, 2904 (1925); h n g s e t h and Niclsen, Phys. Ra., 44,326 (1933); and Brockway and Pauling, Proe. NaII. Acad. Sci., 19, 880 (1933). ( 8 ) Newman, Kuivila and Garrett, THISJOURNAL, 67,704 (1941).

MELVINS. NEWMAN AND HYMAN L. GILDENHORN

318

sibility that dihydroxycarbonium ions6 may also react as shown in equation 2. However, the reaction with dihydroxycarbonium ions is undoubtedly much slower. This difference in reactivity of these two types of carbonium ions is strikingly illustrated by the course of the reaction in the case of I. Findings B and C (above) are also to be noted. The fragments, (RNHCO)+ and [RNHC(OH)z]+, react with water to yield the amine and carbon dioxide. In some cases, at least, carbon dioxide is evolved directly from the sulfuric acid solution. We noticed that the fragment (RNHCO)+ should also be the species first formed when an isocyanate is dissolved in sulfuric acid. When we added phenyl isocyanate to sulfuric acid there was an immediate evolution of carbon dioxide and aniline was isolated from the reaction mixture. The reaction pictured in equation (1) takes place at a temperature determined by the ease with which the acid in question yields appreciable quantities of oxocarbonium ion, R-C+=O. In the case of 2,6-dimethylterephthalic acid, I, and 2,4,&trimethylbenzoic acid, a temperature of 10' or lower is sufficient because of the tendency of these acids to undergo complex ionization8sgin sulfuric acid according to equation (3)."J RCOOH

+ 2H2SO4

R-C+=O

+ HaO+ + 2HSO4-

(3)

But in the case of benzoic acid and many other acids, a temperature of 35 to 50' is required for the Schmidt reaction (equation 2). Acids of this type ionize according to equation 4 in sulfuric acid.lO RCOOH

+ HzSOi 1J RC+(OH2) + HSO4-

(4)

The dihydroxycarbonium ions,s RC+(OH2), thus formed evidently do not react with hydrazoic acid as easily as do the oxocarboniumions,6R-C+=O. The higher reaction temperature required for many Schmidt reactions may be needed to shift the equilibrium shown in equation 5 to the right or to provide a higher temperature for the reaction of the less reactive dihydroxycarbonium ions as shown in equation (2). RCf(OH)2

R-C+=O

+ HsO

(5)

Should it prove possible to produce appreciable concentrations of oxocarbonium ions from acids of the benzoic type at lower temperatures by altering the reaction medium, we would expect reaction with hydrazoic acid. Thus it may be that (9) Raymond A. Craig has shown in unpublished work t h a t acid I has a n I factor of 2.4. Although the exact I factor for this acid could not be predicted its magnitude is in the region expected if one considers t h a t tho i factor of 2,4,6-trimethylbenzoic acid is 4 and t h a t for 2,6-dimethylbenzoic acid is about 3.5. Thus substitution of the 4-methyl group by hydrogen causes a decrease in 1 factor (see Treffers and Hammett, TBISJOURNAL, 69, 1708 (1937)). Further suhstitution of a carboxyl group in the 4-position, as in I, would be expected t o produce a further decrease in the i factor. (10) See Hammett "Physical Organic Chemistry," McGraw-Hill Book Co., Inc., New York, N. Y . , 1940, p. 45-56.

Vol. 70

reaction a t low temperatures of a system with hydrazoic acid to yield nitrogen will prove to be a means of determining the presence of oxocarbonium ions, barring complications involving hydrazoic acid. In the case of Schmidt reactions with aldehydes and ketones a similar mechanism has been used to account for the products formed." The reaction of the hydroxycarbonium ion' formed from aldehydes and ketones in sulfuric acidlo with hydrazoic acid is shown in equation 6. OH

1'

0"

+

[R. > C : N I m J Rz -fi

[

RL-C-N:Rz ' " H

1'

+ NI (6)

It should be pointed out that many of these reactions proceed at low temperatures. The complex ion formed loses a proton to yield an amide. In this connection it should prove interesting to compare the migration aptitudes of various groups in the Schmidt reaction with those obtained by Bachmann in the case of the rearrangement of symmetrical pinacols.12 The qualitative results reported for a few ketones show that the migration aptitudes are roughly what one would expect using data from pinacol rearrangements.", l 2 Further application of the principles outlined above allow reasonable mechanisms to be postulated for the reaction of a variety of other compounds with hydrazoic acid under acid conditions.'8 If we return to the first reaction stage of equation l, it is readily seen that this intermediate, r o l+ R-C: N: N: : :N: , could likewise be obtained

I

J

by the addition of a proton to an acid azide. Therefore, since the Schmidt reaction takes place readily at 35 to 50' and most azides decompose in solvents only at higher temperatures, we were led to predict that the Curtius reaction should be acid catalyzed.'* This prediction was verified by showing that the rate of decomposition of benza(11) See Sanford, Blair, Arroya and Sherk, THISJOURNAL,67, 1941 (1945). for a comparison of mechanisms involving hydrazoic acid and hydroxylamine-0-sulfuric acid. (12) See Wallis in Gilman's "Organic Chemistry," John Wiley aad Sons,Inc., New York, N. Y., 1943, p, 969-970. Dr.P. A. S. Smith, University of Michigan, has informed us t h a t he is studying this problem now. (18) See the entire chapter of ref. 3 above. (14) No mention of acid catalysis in the Curtius reaction is made either in the chapter on this reaction by Smith in "Organic Reactions," Vol. 111, John Wiley and Sons, Inc., New York, N. Y.,1946, p. 337,el scq., or in the chapter by Wallis and Lane, {bid., p. 267. cf reg. Since this article was written, Dr. Smith (see ref. 12) pointed out t h a t Th. Curtius, J . prokl. Chcm., [2] bo, 291 (1894), mentions briefly t h e decomposition of azides in ether containing hydrogen chloride.

Jan., 1948

SCHMIDT

REACTIONS AND

CURTIUS

REARRANGEMENT

319

acids were dissolved ino 100% sulfuric acid and the solutions were cooled a t 0 Equivalent amounts of sodium azide were added to each solution during one hour. From the former no gas was evolved but from the latter a brisk evolution of nitrogen took place. Carbon dioxide was also formed but no carbon monoxide. After stirring an additional half hour the solutions were poured on ice. From the former a negligible amount of aniline was isolated but from the latter 51% of mesidine was obtained. The mesidine was identified as its formyl derivative, m. p. Experimental16 175.5-176.5°.24 We also prepared the 1,3,5-trinitroaddition product of mesidine and found it to 3,5-Dirnethy14-aminobenzoicAcid, 11.-To a stirred benzene form deep reddish-brown needles, m. p. 133.5-134.0°.26 mixture of 4.7 g. of pure 2,6-dimethylterephthalic acid17 Anal. Calcd. for CIbHl&eNd: C, 51.7; €1, 4.6; I'i, in 20 cc. of 100% sulfuric acid and 15 cc. of chloroform a t 16.1. Found: C, 51.7, 51.9; H, 4.8, 5.0; N, 16.1, 25" was added 1.05 g. of sodium azide over a periods of two and ode-half hours. After stirring for another fifteen 16.2. Mesidine from Methyl 2,4,6-Trimethylbenzoate .-To a minutes, the mixture was poured on ice and neutralized with sodium hydroxide. The chloroform layer was sepa- solution of 5.2 g. of methyl 2,4,6-trimethylbenzoate in rated and the aqueous layer acidified with dilute hydro- 40 cc. of concentrated sulfuric acid a t 40" was added chloric acid. The white solid which separated was col- slowly 1.56 g. of sddium azide. The gases evolved lected and dried. It weighed 2.3 g. (87% based on azide amounted to 1020 cc., almost twice the theory for nitrogen. used) and 2elted with darkening and decomposition from No tests on this gas were made other than to show it con240 t o 250 After two recrystallizations from alcohol tained carbon dioxide and did not contain carbon monwhite crystals of pure I1 were obtained. The melting oxide. The reaction mixture was poured on ice and the point of 251 to 252" with decomposition18 was determined mesidine formed was extracted from the alkaline solution in a copper block by heating from 30 to 200" in ten minutes and distilled. I t boiled a t 80" a t 2 mm. and was isolated and then heating a t a rate of 4" per minute. in 79% yield (based on azide). From the acid portion Anal. Calcd. for C Q H ~ ~ O ~C,N :65.4; H , 6.7; N, of the reaction mixture was recovered 13% of 2,4,6-trimethylbenzoic acid. 8.5. Found: C, 64.9, 65.0; H , 6.6, 6.8; N, 8.5, 8.7. Acid Catalyzed Curtius Rearrangements .26-QualitaThis amino acid was diazotized and coupled with 8- tive experiments on the effect of acids on the decomposition naphthol t o yield a brilliant orange-red product, m. p. of benzazide at 75" were carried out in a ~OO-CC., three285.5 to 288' with decomposition, when recrystallized from absolute alcohol (m. p. procedure same as for 11). necked flask fitted with a sealed stirrer and dropping funnel and connected to an azotometer. Various acids were Anal. Calcd. for C I Q H ~ ~ O ~C,N 71.2; ~: H , 5.0; N, added to the acetic acid to make 50 mole per cent. solu8.8. Found: C, 71.0, 71.0; H , 5.2, 5.4; N , 8.8, 8.9. tions. The apparatus was equilibrated a t 75" and was In another experiment, the diazonium solution was then pinched off and the 'flask cooled. A solution of heated with added copper sulfatelotooyield a small amount benzazide in 10 cc. of acetic acid was then sucked in of colorless crystals, m. p. 222-224 , of 3,5-dimethyI-4- through the separatory funnel, followed by a rinse with hydroxybenzoic acidPo on crystallizadon from benzene. 5 cc. of acetic acid. The pinchcock was then removed Further evidence for the structure of I1 was obtained in and the flask heated to 75". The times required for coldecarboxylation experiments. Our amino acid decar- lection of one-half of the total volume of gas evolved boxylated with difficulty whereas NoyesI7reported that the were as follows: acetic acid, 20 minutes; chloroacetic isomeric 2,6-dimethyl-4-aminobenzoicacid decarboxylates acid, 13 minutes; dichloroacetic acid, 10 minutes; trireadily. On heating our acid with a little copper carbo- chloroacetic acid, 7 minutes; 2.5 cc. of 100% sulfuric nate" a small amount of an amine ,was obtained which acid in 50 cc. of acetic acid, 10 minutes ; 50 mole per cent. formed an acetate, m. p. 170-174' after crystallization sulfuric in acetic, 6 minutes. When benzazide was from alcohol and from benzene. The melting point of added to 100% sulfuric acid, an explosion occurred.16 the acetate of sym-m-xylidine is 140.5"*' whereas that of In three experiments in dioxane, the time for dioxane was vic-m-xylidine melts a t 177".** 60 minutes, for a solution of 9 g. of boron trifluoride in 50 Schmidt Reactions a t Low Temperature.-In two paral- cc. of dioxane, 7 minutes, and for 2.3 g. of hydrogen lel experiments 5 g. of benzoic and 2,4,6-trimethylbenzoic chloride in 65 cc. of dioxane, 6 minutes.

zide in acetic acid was increased progressively by separate additions of chloroacetic, dichloroacetic, trichloroacetic and sulfuric acids. The rate of decompositioii of benzazide in dioxane was also markedly increased by hydrogen chloride and by boron trifl~oride.'~We are now studying some quantitative aspects of this catalytic reaction.

.

.

(15) The following references mention explosions or catalysir of decomposition of azides by acids: Bar., 68, 1221 (1925); Ann., 469, 44 (1929); Hclu. Chim. Acfa., 16, 935 (1932), 16, 349 (1933); Rac. Irav. chim., 58, 988 (1934). (16) All melting points corrected. Microanalyses by WI J. Pblglase, 0. S. U. (17) Noyes. Am. Chcm. J . , SO, 789 (1898). (18) Fittig and Brueckner. Ann., 147, 40 (1868), give a m. p. of 235' and Wheeler and Hoffman, THISJOURNAL, 44, 120 (1922), 242O for this compound. (19) Lassar-Cohn "Arbeitsmethoden," Leiprig, 1923, p. 136. (20) Thiele and Eichwede, Ann.. 811, 372 (1900). The isomeric 2,6-dimethyl-4-hydroxybenzoicacid melts at 185', Rabe and Spence, Ann.. 849, 361 (1905). (21) Fieser "Experiments in Organic Chemistry," D. C. Heath Co.,New York, N . Y.,1941, p. 201. (22) Wroblewski, Ann., 907, 96 (1881). (23) Hodgkinson and Limpach, J . Chcm. Soc., 77, 67 (1900).

Summary A mechanism for the Schmidt reactions of acids and certain esters with hydrazoic acid in sulfuric acid solution is proposed. Catalysis of the Curtius rearrangement of benzazide by protonic and non-protonic acids has been demonstrated. COLUMBUS 10, OHIO

RECEIVED AUGUST1, 1947

(24) Hantzsch and Lucas. Bcr., 98, 751 (1895). (25) Noelthing and Sommerhof, {bid., 89, 77 (1906). give 120-122° as the melting point of this derivative. (26) For e study of solvent effect on the rate of Curtius rearrangement of benzazide, see Newman, Lee and Garrett, THISJOURNAL, 69, 113 (1947).